Background: Frozen shoulder is a common disorder that leads to substantial functional loss for patients by impairing activities of daily living. It also adversely affects patients and society by impairing the ability to work. Its pathogenesis is not fully understood. The aim of the present study was to perform a systematic review and meta-analysis to assess the evidence suggesting a genetic link to frozen shoulder.
Methods: A literature search of MEDLINE, EMBASE, and CINAHL databases using relevant keywords revealed 5506 studies. After appropriate screening of titles, abstracts, and full studies, seven studies were analyzed.
Results: Three studies investigated rates of frozen shoulder among relatives. One study (n = 1828 twin pairs) showed an 11.6% prevalence in twin pairs and demonstrated a heritability of 42% for frozen shoulder after adjusting for age. A second study (n = 273) showed that 20% of patients with frozen shoulder had a positive family history involving a first-degree relative. The relative risk of frozen shoulder was 4:1 when all patients with frozen shoulder were compared with a control population. A third study (n = 87) showed that 29% of patients with frozen shoulder had a first-degree relative with frozen shoulder. Two studies evaluated racial predilection for frozen shoulder. One study (n = 50) reported a substantially higher number of white patients (76%) with frozen shoulder than black patients (24%). A second study (n = 87) showed that being born or having parents or grandparents born in the British Isles were risk factors for frozen shoulder. Four immunological studies investigated human leukocyte antigen (HLA)-B27 as a risk factor for frozen shoulder. Meta-analysis of two of these studies with clearly defined controls showed significantly higher rates of HLA-B27 positivity in patients with frozen shoulder as compared with controls (p < 0.001).
Conclusion: The limited evidence points toward a genetic link to frozen shoulder. We used family history and racial predilection as markers for genetic association, both of which indicated the presence of a genetic predisposition to frozen shoulder. However, as there is a lack of unbiased genetic approaches, there is an opportunity for genome-wide association studies to address definitively the molecular genetics of frozen shoulder. Such studies may eventually lead to a better understanding of the pathogenesis of frozen shoulder and the development of novel therapeutic interventions.
Level of Evidence: Prognostic Level IV. See Instructions for Authors for a complete description of levels of evidence.
Frozen shoulder, or adhesive capsulitis, is a disorder that manifests with pain and restricted active and passive motion of the shoulder. The diagnosis of this condition is based on clinical criteria. Frozen shoulder is common, with a reported prevalence of 2% to 5% in the general population1-4. Its prevalence is even higher (10% to 36%) in patients with diabetes5,6. Frozen shoulder most commonly occurs in working-age patients who are forty to sixty years old2,5, with the peak age in the mid-fifties. Women are affected more often than men. The condition is often bilateral, and, in 6% to 17% of patients, it affects one shoulder first and the contralateral side within five years1,2,5,7,8.
The pathogenesis of frozen shoulder is not fully understood. The condition can be either primary or secondary. In cases of primary or idiopathic frozen shoulder, no findings in the clinical history or physical examination explain the onset of disease. In cases of secondary frozen shoulder, the patient may have a history of trauma or surgery, myocardial infarction, diabetes mellitus1,9-11, hypothyroidism12,13, Parkinson disease14, or barbiturate use15.
Frozen shoulder leads to substantial functional loss, with impairment of activities of daily living (dressing, washing, driving, self-care). It also adversely affects both patients and society by impairing the ability to work. Even though frozen shoulder traditionally has been considered to be a self-limited disease, with a natural history lasting two to three years16, it is increasingly recognized that, for many patients, frozen shoulder is a chronic disease. Previous studies have shown that up to 40% of patients have symptoms and restriction of movement that persist beyond three years17 and that 15% have long-term disability18,19. In a Dutch study, the health loss for patients with frozen shoulder was estimated to be 0.048 quality-adjusted life year (QALY) and the societal costs were estimated at €4521 ($4923) per patient, of which 47% were health-care costs and 53% were non-health-care costs20.
The aims of treatment for frozen shoulder are to decrease pain and to improve the range of motion. Current treatments include both noninvasive options (physiotherapy) and invasive options (steroid injections, hydrodilatation manipulation with the patient under anesthesia, or arthroscopic release). Exploring a possible genetic component in frozen shoulder could shed light into its pathogenesis and unveil potential novel targets for medical intervention.
The purpose of the present study was to perform a systematic review and meta-analysis to assess the evidence suggesting a genetic link to frozen shoulder.
Materials and Methods
A systematic literature search of MEDLINE/PubMed (1879 to present), Excerpta Medica Database/EMBASE (1947 to present), and Cumulative Index to Nursing and Allied Health Literature/CINAHL (1961 to present) databases was conducted from their year of inception to May 2014, with the following combinations of keywords: “frozen,” “shoulder” and “adhesive,” “capsulitis,” and “shoulder.” There was no language limit. The words “frozen shoulder genetic” were also entered into a Google search engine, and the first twenty records were evaluated. From the Google search, one additional relevant article was identified. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) methodology guidance was employed21. Full texts were reviewed for relevant articles and in cases in which a decision regarding inclusion could not be made on the basis of the title and abstract. The reference lists of selected articles were also examined for any additional articles that were not identified from the database search. The included articles (apart from the case reports) were appraised critically with use of the revised and validated version of Methodological Index for Non-Randomized Studies (MINORS) and were scored, with the maximum score being 2422. Authors of included studies were contacted if further information was needed. Studies (published in full or abstract form) were included if they examined (1) rates of frozen shoulder among relatives (family history), (2) race and the risk of frozen shoulder, (3) human leukocyte antigen (HLA)-B27 and the risk of frozen shoulder, and (4) other genetic risk factors. Reviews, noncomparable studies, and case reports were excluded. Data were extracted in a standardized manner.
A meta-analysis was performed of studies that investigated HLA-B27-positive rates in patients with frozen shoulder as compared with controls. The meta-analysis was performed with use of a random-effects model. Summary risk ratios and 95% confidence intervals (CIs) were calculated. Heterogeneity was assessed using tau2, I2, Q, and p values. Data were analyzed with Comprehensive Meta-Analysis Version 2 (Biostat). Two studies were excluded as the control groups were not defined and a sample size for the controls was not given. Hence, meta-analysis of the remaining two studies with clearly defined controls was performed.
The search identified 5506 articles by title; thirty-eight initially were selected on the basis of information gathered from the title. Fourteen duplicates were discarded. Twenty-four abstracts were reviewed, resulting in the exclusion of seven articles. A full-text review of the remaining seventeen articles and a thorough search of their references was performed; seven of these articles met the inclusion criteria and were used for analysis (Fig. 1). Demographic data and the level of evidence of these studies are shown in Table I.
Three studies investigated the rates of frozen shoulder among relatives (positive family history)23-25. One of these studies (n = 1828 twin pairs) showed an 11.6% prevalence of frozen shoulder in twin pairs and demonstrated a heritability of 42% for frozen shoulder after adjusting for age (Table II)23. Another study (n = 273) showed that 20% of patients with frozen shoulder had a positive family history involving a first-degree relative (Table II)24. The relative risk for frozen shoulder was 4:1 when those patients were compared with a control population (Table II). A third study (n = 87) showed that 29% of patients with frozen shoulder had a first-degree relative with frozen shoulder, as compared with 7% of individuals in a control group (n = 176) (Table II)25.
Two studies investigated racial predilection for frozen shoulder25,26. One study (n = 50) of patients with frozen shoulder included a substantially higher number of white patients (76%) than black patients (24%) (Table III)26. The other study (n = 87) showed that being born in the British Isles or having parents or grandparents born in the British Isles were risk factors for frozen shoulder (Table III)25.
Four studies investigated HLA-B27-positive rates in patients with frozen shoulder and controls (Table IV)26-29. Meta-analysis of two of these studies with clearly defined controls27,29 showed significantly higher rates of HLA-B27 positivity in patients with frozen shoulder as compared with controls (risk ratio, 3.28 [95% CI, 1.81 to 5.96; p < 0.001]) (heterogeneity: tau2 = 0.068, I2 = 34.3%, Q = 41.52, degrees of freedom = 1, p = 0.22) (Fig. 2).
The critical appraisal of the included studies according to MINORS criteria is summarized in Table V22. None of the studies included consecutive patients, and none was prospective. All of the studies had end points appropriate to their aim and had adequate and baseline-equivalent control and contemporary groups. Three studies had a clearly stated aim23,25,26, two showed evidence of unbiased assessment of the end point23,26, five had follow-up appropriate to the aim of the study23-27, and two reported <5% loss to follow-up26,27. The lowest score was 929, and the highest was 1923.
Frozen shoulder is a common condition that confers substantial morbidity. Its pathogenesis is unclear. The identification of a genetic association could elucidate its underlying pathogenesis and encourage the development of medical treatments. The present systematic review explored studies that evaluated the association of frozen shoulder with family history, race, and HLA-B27 and also explored any other genetic studies examining a genetic component.
Evidence from family studies takes the form of twin studies and evidence of familial clustering. One study showed an overall prevalence of 11.6% of frozen shoulder in 1828 unselected female twin pairs (Table II)23. That study evaluated concordance in monozygotic and dizygotic twin pairs and found that it was higher in the former group (0.30 compared with 0.25) (Table II). That study also calculated an age-adjusted heritability of 42% for frozen shoulder (Table II).
Familial clustering of patients with frozen shoulder has been reported in some studies23,24,30-32. Such clustering may be caused by genetic factors or shared environmental exposure. A positive family history of a first-degree relative was reported in two studies23,24. In addition to the selected analyzed studies, we identified three case reports that did not meet the inclusion criteria for the systematic review and were excluded; each of these reports included two first-degree relatives who developed frozen shoulder30-32. One report described bilateral frozen shoulder in two brothers at the same time30, the second report described frozen shoulder in a woman and her father31, and the third report described frozen shoulder in two identical twins32.
Racial predilection is another factor that points toward a genetic predisposition. Two studies evaluated racial predilection for frozen shoulder23,26 and showed some preference for whites26 or individuals born in or having parents or grandparents born in the British Isles23. This link is similar to that for Dupuytren disease, a higher incidence of which has been linked with Nordic origin33. Kilian et al. evaluated the arthroscopic findings associated with frozen shoulder and noted synovitis followed by a thickened capsule34. This thickened capsule consists of a dense matrix of type-I and type-III collagen laid down by fibroblasts and myofibroblasts. Similar fibrosing changes occur in Dupuytren disease; hence, frozen shoulder has been described as “Dupuytren-like disease.”35 It is of note that a genetic link has been shown in Dupuytren disease and that a relationship between frozen shoulder and Dupuytren pathogenesis has been reported on the basis of histological and immunohistochemical studies35-38. A recent study examining mRNA expression and subsequent collagen I and III synthesis in samples from patients with frozen shoulder and Dupuytren disease provided additional evidence of a common pathogenesis mechanism for the two conditions39. It showed a significant increase of a1 (I) mRNA chains (collagen I) in samples from patients with frozen shoulder (p = 0.016) and Dupuytren disease (p = 0.041) compared with normal capsule, but there was no difference in a1 (III) mRNA chains (collagen III). There was also a decreased rate of expression of intracellular procollagen I and extracellular mature collagen I in samples from patients with frozen shoulder and Dupuytren disease as compared with normal tissue. The increased levels of a1 (I) mRNA chains along with the decreased procollagen I levels suggest that the message was not translated or that the chains were secreted rapidly into the extracellular space. The lack of expression of a1 (III) mRNA chains for collagen III synthesis could be explained by apoptosis of myofibroblasts.
Cytokines and growth factors regulate the growth and overall function of fibroblasts. Remodeling of connective tissue matrix as well as degradation of extracellular matrix (including collagens, elastins, gelatin, proteoglycans) is controlled by matrix metalloproteinases (MMPs), which are secreted by various connective-tissue and pro-inflammatory cells, including fibroblasts, osteoblasts, endothelial cells, and macrophages. Among the collagenases described are MMP-1, 2, 3, 8, 13, and 14. Among the interleukins, IL-1α, IL-1β, and IL-6 are known to mediate inflammatory responses, and IL-8 is a neutrophil chemotactic factor. Transforming growth factor-beta (TGF-β) controls cell proliferation and cell differentiation and induces cell apoptosis. Tumor necrosis factor-alpha (TNF-α) is involved in systemic inflammation. The expression of certain cytokines (IL-1α, IL-1β, IL-8), growth factors (TGF-β), tumor necrosis factors (TNF-α), and MMPs (MMP-1, 2, 3, 14) has been shown to be abnormal in patients with Dupuytren disease40-42. In fact, the expression of all four of these MMPs has been found to be correlated with poor prognosis after fasciectomy in patients with Dupuytren disease43. Bunker et al. investigated the expression of cytokines, growth factors, and MMPs in tissue samples from patients with frozen shoulder44. mRNA levels for some cytokines (IL-1α and IL-6) and growth factors (TNF-α) were higher in samples from patients with frozen shoulder as compared with control tissue. The expression of mRNA for MMP-3 and 9 was also increased. Similarities in the expression of these factors in patients with Dupuytren disease and frozen shoulder suggest a similar molecular pathogenesis mechanism and, hence, possible similar genetic links.
Several studies have investigated possible genes associated with Dupuytren disease. Although a specific gene has not been identified, there are some indications as to where such genetic linkage may be located. A linkage to a single 6 cM region between markers D16S419 and D16S3032 on chromosome 16 was shown in a five-generation Swedish family in which multiple members were affected by Dupuytren disease44. Another study (n = 6) showed that five genes (including those for Alzheimer disease amyloid A4 protein precursor, archain, brain-specific tubulin a1 subunit, protein kinase PKX1, and SEF2-1B protein) were consistently overexpressed in tissue from patients with Dupuytren disease45. A Dutch genome-wide association study tested the thirty-five single-nucleotide polymorphisms (SNPs) most strongly associated with Dupuytren disease and showed an association with eleven SNPs from nine different loci46. More specifically, a locus on chromosome 7 with four markedly associated SNPs was found, with the strongest association at gene EPDR1. Three other associated SNPs were identified at a locus on chromosome 22. The most substantial SNP was on chromosome 22q between genes WNT7B and LOC100271722. One more locus was identified on chromosome 19, with one substantial SNP. It is possible that the same genetic associations occur in frozen shoulder; hence, initial evaluation of these loci may be appropriate when looking for frozen shoulder genes.
Immunological studies have previously examined the relationship of HLA-B27 with the risk for frozen shoulder. HLAs are highly polymorphic proteins. They are encoded on chromosome 6 as part of the human major histocompatibility complex (MHC)47. The MHC is a set of cell-surface proteins that controls a major part of the immune system48. The function of the MHC is to bind peptide fragments derived from pathogens and to display them on the cell surface for recognition by the appropriate T cells49. The MHC contains several different MHC class-I and MHC class-II genes, so that everyone has a set of MHC molecules with different ranges of peptide-binding specificities48,49. MHC class I is found on all nucleated cells and presents epitopes to cytotoxic T lymphocytes (CTLs). A CTL expresses certain receptors (CD8) that bind to an MHC class-I molecule and the CTL triggers the cell to undergo apoptosis. Thus, MHC class I helps to mediate cellular immunity, which is crucial to address intracellular pathogens, such as viruses and some bacteria. HLA is a class-I MHC protein that is expressed on the cell surface of almost all nucleated cells. The B27 type is the most common type of HLA. It has been studied extensively because of its strong association with spondyloarthropathies (e.g., ankylosing spondylitis)50,51. HLA-B27 also has been associated with reactive arthritis, undifferentiated arthritis (arthritis of no specific type), inflammatory bowel disease, and eye disorders (e.g., anterior uveitis, iritis)52-54. Patients with these disorders show an increased frequency of HLA-B27 protein positivity as compared with normal controls. It still is not clear how HLA-B27 influences the process of these diseases, but it is thought that HLA-B27 is involved in triggering an inflammatory response. We are aware of four available relevant studies that have evaluated the association of positive HLA-B27 protein with frozen shoulder, with slightly controversial results (Table IV)26-29. During the meta-analysis of these studies, two were excluded because control groups were not defined and the sample sizes of the control groups were not given26,28. Meta-analysis of the other two studies with clearly defined controls27,29 showed significantly higher rates of positive HLA-B27 in patients with frozen shoulder. These higher rates of HLA-B27 positivity in patients with frozen shoulder suggest a genetic link, but this relative risk and associated link need to be investigated further with bigger studies before the link is established.
The findings of the present study suggest that genetic factors may be important in the etiology of frozen shoulder. However, there is a lack of unbiased genetic approaches to address the etiology of this condition. There is an opportunity for genome-wide association studies to address definitively the molecular genetics of frozen shoulder. Such studies eventually may lead to improved prognostic information, which may alter initial treatment and identify novel therapeutic targets, possibly leading to alternative nonoperative treatments and secondary prevention strategies.
The main limitation of the present study is the definition of frozen shoulder. There is not a unique standardized definition for this condition. The diagnosis is based on clinical criteria, and this diagnosis is not always unbiased. One study24 based the diagnosis on the criteria described by Codman55, two studies based the diagnosis on other similar clinical and radiographic criteria25,27, and four did not clarify the criteria used23,26,28,29.
In conclusion, the limited evidence points toward a genetic link. We used family history and racial predilection as markers for genetic association, both of which indicated that a genetic predisposition to frozen shoulder is present. A large epidemiological study along with a genetic study to identify whether there is a genetic locus for frozen shoulder would be of great value and would shine additional light into the pathogenesis of this condition.
Source of Funding: No external funds were received in support of this study.
Investigation performed at the Orthopaedic Department, Blackpool Victoria Hospital, Blackpool, Lancashire, United Kingdom
Disclosure: None of the authors received payments or services, either directly or indirectly (i.e., via his or her institution), from a third party in support of any aspect of this work. None of the authors, or their institution(s), have had any financial relationship, in the thirty-six months prior to submission of this work, with any entity in the biomedical arena that could be perceived to influence or have the potential to influence what is written in this work. Also, no author has had any other relationships, or has engaged in any other activities, that could be perceived to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitted by authors are always provided with the online version of the article.
- Copyright © 2016 by The Journal of Bone and Joint Surgery, Incorporated